8700k highest IPC to date...leak

Well... yeah?

I mean, nobody was expecting Intel to suddenly kill it's IPC advantage for absolutely no apparent reason in order to push Netburst 2.0.

How is this news?

Ask wccftech im just reporting something else to read about coffee lake.
 
Its already in the CFL thread. And turbo isn't working for MT in that leak.
Yeah those MASSIVE all about threads suck because not everyone is going to read 800 posts in a 12 page thread to see that.

So a fresh posting is just fine and not a big deal.
 
Clock for clock as good as Ryzen in MT, but overall it will do better than the 6-core Ryzen since clocks are higher. Also, very impressive ST results!
 
Clock for clock as good as Ryzen in MT, but overall it will do better than the 6-core Ryzen since clocks are higher. Also, very impressive ST results!

Clock for clock? Core for core? Certainly not. Same reason we see Zen failing to or even below SB IPC more often than not recently ;)
 
Well... yeah?

I mean, nobody was expecting Intel to suddenly kill it's IPC advantage for absolutely no apparent reason in order to push Netburst 2.0.

How is this news?

Agreed. Why wouldn't the newest tech in this category have the highest ipc? Especially from intel?
 
So in other words it's the same IPC as Skylake/Kabylake just with an added bump in clock speed at stock.
 
  • Like
Reactions: N4CR
like this
Yep, but almost a 30% improvement in MT performance over the 7700k, which is very nice.
 
What's MCE? Google isn't any help

(Multi core enhancement).

When mobo makers likes to run CPUs at max turbo for all cores.

Old example:
49379.png
 
A lot more than 30%. The 1210 number is without turbo. Its a ~50% MT increase that shouldn´t be a surprise to anyone. With MCE on both CPUs it will be even more.

Ah, makes sense. I didn't see that. What is the non-turbo freq for all 6 cores on the 8700k?
 
Nevermind, I found it. It is 3.7 ghz base for the 8700k. Impressive, since the 1600x needs 4.0 Ghz to match with 6 cores
 
So in other words it's the same IPC as Skylake/Kabylake just with an added bump in clock speed at stock.

The IPC is not exactly the same because there is more cache and faster RAM. And besides the bump on clocks there is 50% moar cores!
 
How is the Z270 buyer screwed? It still does what it was designed to do.

Because there is no real reason why they couldn't have enabled support. It is still 14nm. It's the same area architecture. They wanted a cash influx from a new chipset. That's their prerogative. And it's mine to choose to wait for a real next gen chipset when I would have gladly spent $350 for a new CPU to put in my 8 month old z270. I'm sure I'm not the only one who feels that way also.
 
Because there is no real reason why they couldn't have enabled support. It is still 14nm. It's the same area architecture. They wanted a cash influx from a new chipset. That's their prerogative. And it's mine to choose to wait for a real next gen chipset when I would have gladly spent $350 for a new CPU to put in my 8 month old z270. I'm sure I'm not the only one who feels that way also.

The same architecture does not mean the same requirements from the chipset to run optimally.

Sometimes you need a different timing on some control lines. Sometimes power delivery requirements change. Sometimes there are bugs in the previous chipset which were relatively harmless with the old CPU but very problematic in the new. All these are things you would not spot looking at the socket, or architecture.
 
This is an speculation and guesswork at best. Prove it.

Intel-Coffee-Lake-S-6-Core-Core-i5-Processor-Specs.png


Note the KBL platform under the "Computer/Device." If they can get ES chips to run on the 200 series chipsets, even with minor adjustments made from ES chips to retail, they could have done it if they wanted. They didn't want to. Why not? They want to sell a new chipset.

The same architecture does not mean the same requirements from the chipset to run optimally.

Sometimes you need a different timing on some control lines. Sometimes power delivery requirements change. Sometimes there are bugs in the previous chipset which were relatively harmless with the old CPU but very problematic in the new. All these are things you would not spot looking at the socket, or architecture.

That might be true, except it is pretty well established at this point that Z270=Z370 with some Intel magic sprinkled in to make them CFL compatible.
 
Note the KBL platform under the "Computer/Device." If they can get ES chips to run on the 200 series chipsets, even with minor adjustments made from ES chips to retail, they could have done it if they wanted. They didn't want to. Why not? They want to sell a new chipset.
How do you know it was not being tested on a modified z270 chipset (which is what z370 is) that is still be reported as such?
 
How do you know it was not being tested on a modified z270 chipset (which is what z370 is) that is still be reported as such?

That's as much speculation as you accused me of ;). Obviously, I have my own feelings on the matter. I will not be buying a 8700k. I sold off my Z270 setup and bought a used X99/5930k one for essentially what a bare 8700k would have cost me. I would have spent the $350 for a 8700k at launch no questions asked if it were available without having to spend another $150-200 for a motherboard.
 
That might be true, except it is pretty well established at this point that Z270=Z370 with some Intel magic sprinkled in to make them CFL compatible.

Isn't Z270 a rebranded Z170 to begin with? kaby lake runs on both Z170 and Z270, maybe it has something to do with that?

I'm sure there are some people around with Sandy bridge era CPU's who are also upset they need to buy a new motherboard for coffee lake.
 
Isn't Z270 a rebranded Z170 to begin with? kaby lake runs on both Z170 and Z270, maybe it has something to do with that?

I'm sure there are some people around with Sandy bridge era CPU's who are also upset they need to buy a new motherboard for coffee lake.

I'm sure that you can notice a difference between a chipset that was released in Q1'17 and one released in Q1'11, but we both know you were only making a snide remark and not being serious.

It's not unlike Intel to offer 3 "generations" on the same platform before. Z87/Z97 had Haswell, Haswell "refresh," and Broadwell support. Let's face it. Kaby Lake could have easily been called Skylake Refresh. Coffee Lake has much more in common with Skylake than it does with Ice Lake, yet the Z370 is supposed to work for both?
 
Yes...but you do not see me making definitive statements on the subject.

Fair enough, but I'm sure we can all be in agreement that it wouldn't have hurt the end user for it to be compatible. Intel stands to gain ~$50 per chipset sold to manufacturers. It helps Intel's bottom line to rebrand a chipset and release it as being a requirement for new CPU's. And up until the ASRock tweet and the CFL announcement with the leaked box it was widely assumed by the community at large (tech websites, forums, even Forbes) that it would be compatible as there is nothing extraordinary about CFL other than the extra cores. It isn't such a large leap of logic to see that Intel believes it has a desirable chip on their hands and they can maximize profits by ignoring a chipset released a few months ago in favor of a rebranded one where they make an extra $50.
 
It's not unlike Intel to offer 3 "generations" on the same platform before. Z87/Z97 had Haswell, Haswell "refresh," and Broadwell support. Let's face it. Kaby Lake could have easily been called Skylake Refresh. Coffee Lake has much more in common with Skylake than it does with Ice Lake, yet the Z370 is supposed to work for both?

Not sure the Z370 will support Icelake, Z390 will, but that one will only arrive 2h of 2018 (even though at first 1q of 2018 was rumoured).

Most likely intel (and their board partners) know that probably a ton of people will run to the new 6 core CPU's and will try and get the most out of it, if coffee lake was only 4c/8t like the older ones they might have kept backwards compatibility.
 
Most likely intel (and their board partners) know that probably a ton of people will run to the new 6 core CPU's and will try and get the most out of it, if coffee lake was only 4c/8t like the older ones they might have kept backwards compatibility.

That's exactly my point. It's a money grab.
 
Back
Top